JP2003148934A - Method for forming dental model from image - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for forming a dental model without requiring aligninment and not requiring bringing a scanning head into a stationary state. SOLUTION: This method for forming the dental model from a series of images of an object in an oral cavity comprises (a) capturing the series of images of the oral cavity of the subject including a common surface characteristic and a control target arranged to the oral cavity so as to give a control characteristic from a plurality of capturing positions; (b) measuring the common characteristic from the images of the object and the control characteristic from the control target imaged together with the images of the object; (c) aligning the measurement result of the control characteristic by photogrammetry to analytically generate the three-dimensional model of the object, reducing the error of the images by the variable directions of the capturing position, and giving the photogrametrically aligned threedimensional model of the object; and (d) adjusting the photogrametrically alined three-dimensional model of the object by aligning the common characteristic of the model to the similar characteristic on the image of the object to generate the aligned dental model from the series of images.

Description

Detailed Description of the Invention

[0001]

FIELD OF THE INVENTION The present invention relates generally to the field of tooth imaging, and more particularly for controlling the automated manufacture of dental inlays for imaging a tooth prepared cavity and subsequent cavity entry. And method for automatically generating a model.

[0002]

BACKGROUND OF THE INVENTION The present invention relates generally to the following conventional situations. The dentist prepares a cavity in the cavities to allow restoration of the tooth, for example by an inlay or crown. After preparation is made, the cavity is cast and typically sent to a dental lab. Unlike such conventional methods, there is a different state-of-the-art technique for reducing the role of the dental lab and producing desired restoration labs. In particular, the prepared cavity is recorded using an electro-optical scanning head. The data thus obtained is supplemented by operator input using techniques in the field of CAD (Computer Aided Design) and the final artifact is manufactured using a small NC (Numerical Control) polisher. .

US Pat. No. 4,837,732 (Brande
stini) describes how a dentist records the raw shape of a tooth prepared for restoration. This method involves the acquisition of data defining the three-dimensional shape of the prepared teeth and their immediate immediate area. First, the raw image from the scan head is displayed on the video display, and while observing the tooth image on the video display, the scan head is manually oriented with respect to the prepared tooth. Then, from the data produced by the scanning head in the selected orientation, corresponding depth and contrast images are generated, which depth images are processed on the basis of the contrast images. The method further superimposes a graphical marker on the image displayed on the video display to facilitate online alignment of the teeth displayed on the raw image with the reference data from previous data acquisitions. Including the step of doing.

[0004]

A drawback of this prior art method is that it involves an alignment method which can later interfere with the quality of the result, and at a particular point in the procedure the dentist can scan heads. Is required to be held almost completely stationary. In particular, generally three-dimensional (3
D) Regarding the artifacts due to the registration method (eg fringes, spots and / or Venetian blind effect), the patents mentioned above are “intolerable and excluded because the phase angle difference is used for depth measurement. Must be done. ” Further, the above-mentioned patent requires that "near-instantaneous 3D acquisition after triggering" is performed, and that the orientation of the scan head must not change between the search mode and the acquisition mode. It is said that it is a general condition.

In the patent by Brandestini et al., The 3D result is overlaid on the search image to allow the dentist to see the result. However, what is needed is a system in which 3D artifacts are projected onto an image using photogrammetric projection equations. This allows the result to appear as if it were actually present in the scene at the time of image capture, allowing for a more accurate and precise evaluation.

However, conventional photogrammetric approaches (see, eg, US Pat. No. 5,372,502) have not worked very well for a number of reasons. For example, such an approach did not work because it was difficult to determine the exact relationship between the camera and the object. In addition, because the teeth are fairly uniform in color and have only a slight texture (which is also partly the reason why the relationships are difficult to determine, as described above) It is difficult.

[0007]

The present invention is directed to overcoming one or more of the problems set forth above. In summary, according to one aspect of the invention, a method (and system) for creating a model of a tooth from a series of images of an intraoral object includes (a) intraoral to provide common surface and control features. Capturing a series of images of an intraoral target from a plurality of capture positions, including a control target placed on the target; and (b) measuring control features from the images of the control target included in the target image. And (c) aligning the control feature measurement results by photogrammetry to analytically generate a three-dimensional model of the object, thereby reducing image errors due to variable orientation of the capture position and the object photo. Providing a 3D model aligned by the survey method, and (d) adjusting the 3D model aligned by the photogrammetric method of the object by aligning common features of the model with similar features on the image of the object. And, thereby and generating a model of the teeth are aligned from a series of images. In practice, the last step involves the application of a 3D morphing algorithm to correct the shift.

[0008]

BRIEF DESCRIPTION OF THE DRAWINGS For the above and other aspects, objects, features and advantages of the present invention, read the following detailed description of the preferred embodiments and claims, and the accompanying drawings. It will be better understood and appreciated by reference.

Since tooth imaging systems and methods are well known, the present invention will be described with particular reference to elements that are part of, or more directly incorporated into, the apparatus and methods of the present invention. Elements not specifically shown or described may be known from the prior art. Some aspects of the embodiments described below may be provided as software. Given the system shown in the drawings and described in the text below, the software not specifically shown, described, or suggested to implement the present invention is conventional and conventional in the art. Within the range of. This is especially true given the current state of the art in the automation of conventional photogrammetry and known photogrammetry.

Referring first to FIG. 4, a preferred embodiment of the present invention includes an intraoral camera 2, a computer system 3 containing instructions for practicing the present invention, and a working tool 4. In the arrangement shown in FIG. 4, the interconnections between the camera 2, the computer system 3 and the working tool 4 are indicated by arrows and are therefore not specifically shown. These interconnections may take various forms and may include, for example, a cable or any other electromagnetic connection (eg, rf transmission), or manual transfer of data from machine to machine. Camera 2
Can be any conventional type of dental camera capable of capturing a fairly high resolution image of an intraoral object, eg tooth 4, a desirable example being incorporated herein by reference, JP Spoonhower, JR Squilla. And commonly assigned co-pending US patent application No. 0 entitled "Intra-Oral Camera with Integral Display" filed February 28, 2001 by JT Boland
It is an intraoral camera described in 9 / 79,239.

The camera 2 is manually supported by the dentist and several images of the teeth are captured. However, it is understood that the orientation of the camera with respect to the teeth will vary from image to image.
One feature of the invention is that it eliminates the effect of different orientations on subsequent measurements. The digitized data from camera 2 is transferred to computer system 3 for processing. The method of the present invention is implemented by the computer system 3 in its processor 5 and the result of the imaging can be displayed interactively on the monitor 6. The operator can operate the cursor 9 with the keyboard 7 and / or the mouse 8 to perform the types of measurements described below. The output from the computer system 3 is a digitized three-dimensional surface pattern which is transferred to the machining tool 4 as a tool path program for the manufacture of tooth molds or dental restorations. The program instructs the milling machine 10 to mill a tooth mold or restoration work piece 11 from a suitable substrate of, for example, ceramic or any other suitable workable material.

[0012] The tooth imaging method according to the present invention is a three-dimensional morphin that causes projection by photogrammetry, analytical adjustment for control, and appearance of an accurate tooth model.
Use the survey method with g). In the present application, the survey method means (1) a step of identifying control points on a digitized image, (2) a step of stereoscopically transferring these points to overlapping images when these points appear, (3 ) Refers to the measurement process including the actual measurement of the image coordinates of the control points.

Photogrammetry is generally a technique for making graphical measurements with light, and more specifically, reliable measurement results by photography or other forms of imaging, such as electronic sensing by sensors. (In general, Manual of Photogrammetry, Fourth Edition,
See American Society of Photogrammetry, 1980). Projection in photogrammetry refers to image projection using an analytical representation of a physical model that describes the imaging process of the sensor. The term projection is specifically projected from the intraoral object through the lens of the sensor to the image plane, which in this case uses the physical model of the imaging process to determine where the points are located. Refers to the concept of light rays.

Analytical adjustment for control refers to the process of correcting a set of parameters representing a physical model into a known or controlled subset of parameters. In general, the least squares adjustment process is applied to a set of linearized conditional equations to a normal set of equations, and the set of linearized conditional equations is the image for the entire parameter set. The partial derivative of the coordinates. Details of least squares processing are well known to those skilled in the art and are described, for example, in Manual of Photogrammetry, Fourth Edi, which is incorporated herein by reference.
tion, op. cit., pp. 77-88.

Three-dimensional morphing is the process of adjusting a three-dimensional target model to a target image. this is,
Project a hypothetical three-dimensional model of the object onto an existing image (through the analytical physical model described above) and detect the deviation between the true image and the image derived from the projected object model. , By correcting the target model (which is then reprojected) to improve the fit. Techniques for three-dimensional morphing are well known in the art and will not be discussed in detail here. For more information, see Frederic Pighin et al. "Systhesizing Realistic.
Facial Expressions from Photographs ", in Computer
Graphics Proceedings, Annual Conference Series, 1
998, pp. 75-83, and Takaaki Akimoto et al., "Aut
omatic Creation of 3D Facial Models ", in IEEE Comp
uter Graphics & Applications, September 1993, pp. 1
See 6-22. Although these publications specifically describe the model of the face, the application of the technique to the model of the tooth is similar.

Referring to FIGS. 1 and 2, a method according to the present invention is shown in which multiple images 12 of an object (one or more teeth 14) in the oral cavity are taken by the camera 2 in several different aspects and / or Or it is captured first from the location. For each image, one or more teeth 14 (eg, teeth 14a) include a control target 16 as shown in FIG. (In practice, the tooth is generally either an unprepared tooth, or a tooth (ie, the root of the tooth) prepared for a restoration procedure.) The target 16 is a saddle-shaped hard tooth. It is a material, and the length C thereof is covered on the tooth 14a so as to extend along the side surface of the tooth (FIG.
Although not shown as such, the control target 16 can span several teeth, such as both teeth 14a and 14b. The lengths A, B, C and D are known and may be unequal lengths. The angle formed by the apex 18 is also known and is a known or controlled parameter described above that is used in the analytical adjustment for the apex 18 and control as described above, as described below. Some goals may be configured to be dimensionally variable to accommodate differently sized teeth. In general, several images were obtained from several different sides / positions as the basis for a three-dimensional view of the intraoral object, and the natural nature of the tooth (or the root of the tooth if the oral object is a prepared tooth). It includes both some common control features and common features of the tooth, such as peaks and valleys, which exhibit a smooth topographical surface.

The surveying process involves measuring common features or parameters (peaks and valleys) in the feature measuring step 20 and measuring control parameters or parameters (apex 18) of the target 16 in the control measuring step 22. Including. There are several ways to make these measurements. Referring to FIG. 4, these measurements indicate that when multiple images 12 are displayed on the monitor 6, the operator positions the cursor 9 over each feature on each image 12, thereby causing each measurement. It can be done interactively by the resulting coordinates being captured by the processor 5. Alternatively, the processor 5 may use any suitable conventional image processing algorithm to automatically locate each feature. This may include image enhancement algorithms and other feature enhancement algorithms as needed. The measurement result is the photogrammetry method adjustment step 24.
In, it is processed to calculate the object space coordinates of any object point imaged in multiple overlapping images from different camera orientations. This process is based on the Manual of Photog
Use the least squares process described above in rammetry, Fourth Edition, op. cit. Basically, this is a multiray stereo intersection process used to find the position of each image point with respect to the camera. The result is a three-dimensional model 26 of the processed tooth with an analytical representation of the physical model representing the imaging process of the sensor that captured the image.

However, the imager is usually a hand-held camera that does not give a perfect geometrical representation of the target, but makes errors due to the lack of certainty of knowledge about the image position. One of the features of the present invention is to address the problem of eliminating these errors, namely variability with respect to camera orientation, before attempting to correct for errors in the actual model of the tooth. In this way, the requirements (and problems) required in the prior art, namely that the scanning head or the imager have to be held completely stationary, can be avoided. Therefore, the analytical adjustment using the control points (vertex 18) to correct the estimate projects the 3D model 26 onto the existing image (one of the multiple images 12) in the analytical projection stage 28. By determining the deviation of the control points (vertex 18) between the model and the image in the deviation step and if the deviation is not acceptable (decision 34), by improving the adjustment by photogrammetry in the improvement step 32. Be seen. Projection is meant to be an analytic process, i.e. to be accomplished mathematically, and the deviation determination can be accomplished interactively (using the cursor 9) or automatically by a suitable image processing algorithm. This projection process
It will be helpful to understand that it uses a physical model to represent the imaging process and is therefore different than simply overlaying the 3D model on the image. With this correction, the variability of the model caused by the various camera positions is reduced or eliminated to an acceptable level.

Once the controlled alignment is acceptable, the slopes and curves between the tips and valleys in the model should or should be matched with corresponding features in the tooth image. Therefore, in the shift determination step 36, determining the residual shift of the model with respect to the actual image, i.e. the shift of the common feature (if any) in the model for the same feature in the actual image. is necessary. If there is a discrepancy (decision 38), the three-dimensional morphing step 40 is initiated to adjust the three-dimensional object model to the image of the model in the adjusting step 42. (This changes the three-dimensional position of the points in the model without the traditional adjustment of camera orientation shifts). This is accomplished by projecting a hypothetical three-dimensional model of the object (through the analytical physical model described above) onto one of the existing images of the intraoral object in the projection step 44. Then, in the shift step 35, the shift between the true image and the image derived from the projected object model is detected. If the deviation is acceptable (decision 48), the process ends. Otherwise, a correction is made (and then reprojected) to the target model in steps 40 and 42 to improve the fit.

At this point, an acceptable model of the tooth has been generated and used for subsequent processing, such as by using the processing tool 4 in either the laboratory or the dentist's office, to produce the desired restoration work. sell. In accordance with the present invention, models of other dental prostheses can be created in addition to teeth, including but not limited to bridges, overlays, and other dental restoration work. Be understood. Further, in addition to milling or cutting, other manufacturing methods such as, but not limited to, injection molding may be used.

FIG. 3 illustrates another method of adjusting a three-dimensional object model to an image of an intraoral object, namely element 40 of FIG.
FIG. 9 is a diagram showing another method logically represented by -48. More specifically, in the alternative method shown in FIG. 3, the generic 3D model from database 50 of such items may be used in addition to or in place of the 3D model provided in elements 40-48. This approach is particularly accurate because it allows the use of a comprehensive tooth model instead, as the tips and valleys may not be well defined by the target vertices. To deal with the problem of difficulty.
This essentially means that in FIG. 1 the 3D model 26 is eliminated or, more precisely, only creates the target 3D model (rather than the tooth and target 3D model), and proceeds from there. It means that it can be reduced to just one thing.

[0022]

The main advantage of the present invention is that by using the projection method and adjustment of photogrammetry for control, for example,
The point is that the projection of the light beam onto the target used in the patents outside Brandestini eliminates the need for alignment schemes that could cause unacceptable artifacts. Furthermore,
In the present invention, no phase information is used, so image capture need not be limited to "almost instantaneous" conditions.

Since the use of the target gives some measure of measurement and allows the determination of the relationship between the camera and the target, it is possible to measure the exact relationship between the camera and the intraoral object. Becomes By using 3D morphing to project the data available on the image,
We deal with measuring an object accurately by letting you see how well it fits in the dimensional object space. If correct, it should "fit" and "fit" the tooth model.

[Brief description of drawings]

1 is a block diagram illustrating a method of creating a tooth model from an image according to the present invention.

2 is a perspective view showing targets useful in the method shown in FIG. 1. FIG.

FIG. 3 is a block diagram illustrating a morphing technique using a comprehensive tooth model database.

FIG. 4 shows a dental system using the method shown in FIG.

[Explanation of symbols]

2 camera 3 computer system 4 processing tools 5 processors 6 monitors 7 keyboard 8 mice 9 cursor 10 milling cutter 11 Repair work 12 images 14 teeth 14a teeth 14b teeth 16 Control target 18 vertices 20 Characteristic measurement stage 22 Control measurement stage 24 Photogrammetry stage 26 3D model 28 Projection stage 30 shift stage 32 Improvement stage 34 decision 36 out of phase 38 decision 40 3D Morphing Stage 42 Adjustment stage 44 Projection stage 46 shift stage 48 decision 50 3D model database

Continued front page    (72) Inventor John P. Spoon Hower             United States New York 14580             Webster Hardwood Lane             1245 (72) Inventor John Earl Squirra             United States New York 14625             Rochester Rutherfield Lane               14 F term (reference) 2F065 AA04 AA51 AA53 CC16 FF04                       JJ08 JJ26 QQ18 QQ24 QQ31                       QQ38 SS01                 5B046 AA00 EA09 FA18                 5B057 AA07 BA02 DA20 DB03 DB05                       DB09 DC09 DC19 DC36

Claims (4)

[Claims] 1. A method of creating a tooth model from a series of images of an intraoral object, comprising: (a) providing a control target placed on the intraoral object to provide common surface features and control features. Capturing a series of images of the target in the oral cavity from a plurality of capture positions; (b) measuring control features from the images of the control target included in the target images; and (c) measuring the control features. Analyzing the three-dimensional model of the object by aligning the results with the photogrammetric method, thereby reducing image errors due to the variable orientation of the capture position and aligning the object with the photogrammetric method. Providing a dimensional model, and (d) adjusting the photogrammetrically aligned three-dimensional model of the object by aligning common features of the model with similar features on the image of the object. Whereby the method comprising the steps of generating a model of the teeth are aligned from the series of images. 2. The method of claim 1, wherein step (b) further comprises measuring the common feature from a series of images of the object. 3. The step (c) comprises the steps of adjusting by photogrammetry, projecting a three-dimensional model of the image by photogrammetry, determining deviations of the control features and correcting the deviations. Thereby improving the photogrammetric adjustment, thereby producing a three-dimensional model of the photogrammetrically aligned object. 4. The step (d) determines the displacement of the three-dimensional model aligned by the photogrammetric method with respect to the image of the object by projecting the model onto the image of the object by the photogrammetric method. The method of claim 1, comprising the steps of: applying a three-dimensional morphing algorithm to correct the offset.